A answering service that picks up the phone but doesn't book a job is an expensive answering service. The headline number that matters isn't the monthly fee. It isn't the per-minute rate. It isn't the count of calls answered. The number that pays the bills is the cost per booked call — the share of the monthly bill you can attribute to a real, scheduled service window.
This page reframes phone-answering spend around the only metric that produces revenue, walks through the calculation with a realistic example, and shows where the levers move.
The first mistake most owners make is dividing the monthly bill by the count of answered calls. That gives you cost per answered call, which is a vanity metric. The metric that drives the business is the cost per booked call, because a booked call is what generates a tech visit, an invoice, and revenue.
A $297/month plan that answers 100 calls and books 30 jobs costs $9.90 per booked call.
A $250/month plan that answers 80 calls and books 6 jobs costs $41.67 per booked call.
The first plan costs more. The second plan is the more expensive option by a wide margin, because most of the "answered" calls didn't go anywhere. They were captured as messages that someone had to call back, or they were answered by a script that didn't know what to do with a broken-spring call, or the caller hung up before a window was offered.
The math is the math. The plan that books more jobs wins, even when it costs more in the invoice. For the formula in detail, see how to calculate your cost per booked call.
The cost per booked call calculation is two inputs:
Cost per booked call = monthly cost of the service ÷ number of jobs booked from those calls
Both inputs are easy to pull. The monthly cost is on the invoice. The number of jobs booked is the count of scheduled service windows that originated from calls handled by the service. Most AI answering providers (Ava included) can give you a booked-call count from their call logs; cross-check it against your own dispatch board.
Once you have the two numbers, divide. The result is what each booked call actually cost the shop. Compare that number to your average ticket to see if the math works.
For a worked walk-through, see booked call vs answered call: know the difference. The difference is where most answering-service bills go to die.
These are planning figures. Substitute your own.
Shop profile: - Two trucks, steady residential work - Monthly answering-service cost: $297 (Ava flat fee) - Calls answered by the service in the month: 180 - Jobs booked from those calls: 45 (most of which the shop would have missed without coverage) - Average ticket: $425
Cost per booked call: - $297 ÷ 45 = $6.60 per booked call
Compare to average ticket: - $425 average ticket vs. $6.60 cost per booked call = 64× ratio
A 64× ratio between the average ticket and the cost per booked call is healthy. The shop is paying less than two percent of one job's revenue to capture the call that books it. Every booked call is margin.
Now run the same math on a per-minute live service that took messages instead of booking:
The per-minute plan costs more and books fewer jobs. The cost per booked call is more than 10× higher. Even though both services "answered" the same number of calls, only one of them produced revenue.
For the levers that move the number, see lowering your cost per booked call.
There's no industry-wide number to copy. The right target depends on your average ticket and your overhead. A reasonable planning framework:
A $6.60 cost per booked call against a $425 ticket is the first category. A $70 cost per booked call against a $425 ticket is the third.
A few real-world factors move the number in ways that aren't obvious from a rate card:
For the broader case for booking over message-taking, see how an AI receptionist books service calls straight into your schedule.
The invoice tells you what the service cost. The cost per booked call tells you whether the service is working. Two plans can have similar monthly costs and very different cost-per-booked-call numbers, because one of them books jobs and the other one takes messages.
A shop owner who only watches the monthly invoice will switch services when the invoice goes up. A shop owner who watches the cost per booked call will switch when the metric drifts, which usually happens months before the invoice does.
The metric also catches plans that look cheap on a rate card but quietly underperform. A $79/month plan that books 4 jobs a month is a $19.75 cost per booked call. A $297/month plan that books 45 jobs a month is a $6.60 cost per booked call. The first plan is cheaper on the invoice and more expensive on the metric that matters.
A side-by-side with the same shop on two different plans, same call volume, same call mix. The difference is what each plan does on the call.
Shop profile: - Two trucks, mixed residential work - Monthly calls: 180 - Average call length: 4 minutes - Close rate on a fully booked call: 50% - Close rate on a message-only call (caller is asked to wait for a callback): 20%
Plan A: per-minute live answering service - Monthly cost: ~$1,100 (base + 720 minutes × $1.50 + small after-hours premium) - Calls "answered": 180 - Calls actually booked: 36 (message-only service at 20% callback close) - Cost per booked call: $1,100 ÷ 36 = $30.56
Plan B: flat-rate AI receptionist (Ava, $297/mo) - Monthly cost: $297 - Calls answered: 180 - Calls booked: 90 (booking on the call, 50% close) - Cost per booked call: $297 ÷ 90 = $3.30
Plan A is more than nine times more expensive per booked call, even though it answered the same number of calls. The headline invoice says one thing; the metric says another. The cost per booked call is the metric that matches the way the shop actually makes money.
The calculation is simple. The discipline is doing it every month.
The first month of any new service is the baseline. Month two is when the trend starts to matter. By month three, you know whether the plan is working and whether it's worth keeping.
The metric points to a switch in two situations:
The flip side: if the cost per booked call is low and stable, the plan is working. Don't switch for the sake of switching — onboarding a new service is a real cost in setup time and call quality variance.
Cost per booked call is one of three numbers worth watching every month. The other two:
The three numbers together tell you whether the marketing is producing real calls, the answering is converting those calls, and the bookings are actually happening. Any one of them by itself is incomplete.
The cost per booked call is the number that decides whether your answering service earns its fee. The monthly invoice tells you what you spent; the metric tells you what you got. A flat-fee plan that books 30+ jobs a month at a $6–$10 cost per booked call is doing its job. A per-minute plan that books 10 jobs a month at $70+ per booking is the more expensive option, no matter what the rate card says.
Ava's flat-fee structure is built to make this metric work in your favor: $297 a month, unlimited calls, every caller captured and booked, with a 30-day "First 10 Leads On Us" guarantee to back it up. If the first 10 leads don't show, the money comes back.
Pull your last month's numbers. Run the calculation. Then have Ava call you and find out what your actual cost per booked call could be.
Call the live demo and have Ava call you now — hear exactly what your customers will hear when they call your shop.